Sigrid Aslaksen
Functional characterization of genetic risk factors in
autoimmune Addison’s disease
2020
Thesis for the degree of Philosophiae Doctor (PhD) University of Bergen, Norway
at the University of Bergen
Avhandling for graden philosophiae doctor (ph.d ) ved Universitetet i Bergen
.
2017
Dato for disputas: 1111
Sigrid Aslaksen
Functional characterization of genetic risk factors in autoimmune Addison’s
disease
Thesis for the degree of Philosophiae Doctor (PhD)
Date of defense: 23.10.2020
The material in this publication is covered by the provisions of the Copyright Act.
Print: Skipnes Kommunikasjon / University of Bergen Name: Sigrid Aslaksen
Title: Functional characterization of genetic risk factors in autoimmune Addison’s disease Year: 2020
Scientific environment
The work of this thesis was conducted between 2017 and 2020 at the Department of Clinical Science and KG Jebsen Senter for Autoimmune Sykdommer, Faculty of Medicine, University of Bergen, with Dr. Eirik Bratland as supervisor and Professor Eystein S. Husebye as co-supervisor. Financial support was provided by the Research Council of Norway (grant no. 262677), and travel grants from Dr. Nils Henrichsen og hustru Anna Henrichsens legat, Det alminnelige medisinske forskningsfond and Scandinavian Society of Immunology.
Acknowledgements
First of all, I sincerely thank my principal supervisor Dr. Eirik Bratland for the continuous support, scientific guidance and for always being available throughout the PhD. Your theoretical knowledge and practical skills have been invaluable during these three years. I am also grateful for all our discussions and brainstorming, which have been so inspiring and encouraging. This work would not have been possible without your supervision and involvement.
I am greatly thankful to my co-supervisor Professor Eystein S. Husebye for offering me the opportunity to become part of the exciting scientific environment within the Endocrine Medicine group. Thank you for sharing theoretical expertise in endocrinology and for supporting this project.
I would also like to express my gratitude to all my past and present colleagues in the Endocrine Medicine group for contributing to such a stimulating and social environment. A special thanks goes to Lars Breivik, Alexander Hellesen, Anette Bøe Wolff, Bergithe Eikeland Oftedal (especially during my visit in Oxford), Amund Berger, Ellen Røyrvik, Åse Bjorvatn Sævik, Elisabeth Halvorsen, Hajirah Muneer, Marie Karlsen, Elin Theodorsen, Øyvind Bruserud, Obaidur Rahman, Solveig Henriette Einevoll, Thea Sjøgren, Shahinul Islam, Andre Sulen, Marianne Øksnes, Elinor Vogt, Paal Methlie and Haydee Artaza Alvarez for valuable discussions, supportive moments, technical support, and for many appreciated lunch breaks.
My sincere thanks to my girls Kristin, Regine, Kristina and Martha for all the good times shared on and off work. Our coffee and lunch breaks, online quizzes during the time of corona, social events, not to mention your good sense of humor, have been especially important to me during the PhD. Thanks to Olivera Bozickvic for being such a positive, kind and motivating person. Thanks to Thomas Helland for great coffee breaks and DJ tips, in addition to the rest of the Hormone group for creating a good working environment. Thanks to Brith Bergum for technical support with flow cytometry analyses.
I sincerely thank my precious family and friends, especially my parents Kathinka and Per for always believe in me and for giving me endless motivation and support. I am also thankful to Aslak and Kari for all the pleasant evenings and discussions at Møllendal Fetevare.
Finally, I want to thank my Per Gunnar for your love, understanding and encouragement during the PhD. Your support in the final stages of my thesis writing has been indispensable.
Bergen, July 2020
Sigrid Aslaksen
Summary
Autoimmune diseases occur when the immune system attacks and damages the body’s own tissue. Why people develop these diseases, and how the autoimmune reaction develops are unanswered questions. Autoimmune Addison's disease (AAD) is an organ-specific autoimmune disorder characterized by an immunological attack of the adrenal cortex. The complex genetic architecture underlying AAD has not been entirely established, and the overall aim of this project was therefore to identify and functionally characterize genetic risk factors in AAD.
We discovered several rare and damaging inborn errors of antiviral immunity in AAD patients. Among them, variants in the gene encoding Toll-like receptor 3 (TLR3), which recognizes double-stranded RNAs (dsRNAs) upon viral infection. Functional characterization of the TLR3 variants revealed a partial loss of function effect on the receptor’s signaling activity, leading to impaired interferon (IFN) responses ex vivo.
Next, we identified a homozygous stop-gain variant in the gene encoding 3β- hydroxysteroid dehydrogenase type 2 (3βHSD2), causing a rare form of congenital adrenal hyperplasia (CAH). The mutation was carried by an AAD patient with circulating antibodies against the major AAD autoantigen 21-hydroxylase (21OH). To our knowledge, this combination represents a novel disease etiology.
Finally, we wanted to identify HLA-specific immunodominant epitopes of 21OH, targeted by autoreactive T cells. We discovered a new immunodominant epitope, ARLELFVVL (21OH434-442), presented by HLA-C*0701. This is the first HLA- C*0701 restricted epitope described for a self-antigen in an autoimmune disease. We also confirmed the presence of autoreactive CD8+ T cells responses to the previously proposed epitope LLNATIAEV (21OH342-350), restricted to HLA-A2.
Altogether, the work in this doctoral dissertation has provided new insights into why certain individuals might be more genetically susceptible to develop AAD, and partly how the autoimmune reaction progresses.
Table of contents
SCIENTIFIC ENVIRONMENT ... 3
ACKNOWLEDGEMENTS ... 4
SUMMARY ... 7
TABLE OF CONTENTS ... 8
ABBREVIATIONS ... 10
LIST OF PUBLICATIONS ... 14
INTRODUCTION ... 15
THE INNATE AND ADAPTIVE IMMUNE SYSTEM ... 15
The cellular components of innate immunity ... 16
Innate immune receptors ... 17
The adaptive immune system ... 19
IMMUNOLOGICAL TOLERANCE ... 22
Central tolerance ... 23
Peripheral tolerance ... 24
AUTOIMMUNITY ... 25
Genetic factors ... 25
Environmental factors ... 27
AUTOIMMUNE ADDISON’S DISEASE ... 29
The adrenal cortex ... 29
Clinical characteristics ... 31
The genetic basis of AAD ... 32
Immunopathogenesis ... 35
Viruses as triggers of the autoreactive immunity ... 37
AIMS ... 41
METHODOLOGY... 43
PATIENT AND CONTROL MATERIAL ... 43
REPORTER GENE ASSAY (PAPER I) ... 43
MHC CLASS I/PEPTIDE MULTIMERS (PAPER III) ... 44
HLA-C TYPING (PAPER III) ... 44
RESULTS ... 45
DISCUSSION ... 51
INVOLVEMENT OF INBORN ERRORS OF INNATE IMMUNITY ... 51
IDENTIFICATION OF AN EXTREMELY RARE HOMOZYGOUS MUTATION IN THE HSD3B2 GENE ... 54
IMMUNODOMINANT HLA-RESTRICTED EPITOPES OF 21OH IN AAD ... 56
CONCLUSIONS ... 61
FUTURE PERSPECTIVES ... 63
REFERENCES ... 65
APPENDIX ... 76
Abbreviations
AAD Autoimmune Addison’s disease ACTH Adrenocorticotropic hormone AH Ancestral haplotype
AIH Autoimmune hepatitis AIRE Autoimmune regulator AITD Autoimmune thyroid disease ALRs AIM2-like receptors
APECED Autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy APC Antigen-presenting cell
APS Autoimmune polyendocrine syndrome BACH2 BTB domain and CNC homolog 2 BCR B cell receptor
CADD Combined Annotation Dependent Depletion CAH Congenital adrenal hyperplasia
CD Cluster of differentiation CD Celiac disease
CLEC16A C-Type Lectin Domain Containing 16A CLRs C-type lectin receptors
CMV Cytomegalovirus CNS Central nervous system CT Computer tomography
CTLA-4 Cytotoxic T lymphocyte-associated protein 4 CVID Common variable immunodeficiency DC Dendritic cell
DHEA Dehydroepiandrosterone DHEA-S DHEA-sulfate
DNA Deoxyribonucleic acid dsRNA Double-stranded RNA EBV Epstein-Barr virus EV Enteroviruses
GWAS Genome wide association study gnomAD Genome aggregation database HCV Hepatitis C virus
HIV Human immunodeficiency virus HLA Human leukocyte antigen
HSE Herpes simplex virus encephalitis HPA Hypothalamic-pituitary-adrenal HSV Herpes simplex virus
IFN Interferon Ig Immunoglobulin
IKKε Inhibitor of nuclear factor kappa B kinase subunit epsilon
IL Interleukin
ILC Innate lymphoid cells
IPEX Immunodysregulation, polyendocrinopathy, enteropathy, X-linked IRAK4 IL-1R-associated kinase 4
IRF3 IFN regulatory factor 3 ISGs IFN-stimulated genes
JAK/STAT Janus kinase/signal transducers and activators of transcription KIRs Killer cell Ig-like receptors
LC-MS/MS Liquid chromatography-tandem mass spectroscopy
LKM-1 Liver/kidney microsomal antibody type 1 LPS Lipopolysaccharides
MAPK Mitogen-activated kinase
MDA5 Melanoma differentiation-associated gene 5 MHC Major histocompatibility complex
MR Magnetic resonance MS Multiple sclerosis
mTECs Medullary thymic epithelial cells
MyD88 Myeloid differentiation primary response 88 NALP1 NACHT leucine-rich-repeat protein 1 NF-kB Nuclear factor-kappa B
NK Natural killer NLRs NOD-like receptors
NOD Nucleotide-binding oligomerization domain PAMP Pathogen-associated molecular patterns PBMCs Peripheral blood mononuclear cells PD1 Programmed cell death protein 1 PGN Peptidoglycan
PIDs Primary immunodeficiency diseases POI Premature ovarian insufficiency Poly(I:C) Polyinosinic:polycytidylic acid PRRs Pattern recognition receptors
PTPN22 Protein tyrosine phosphatase non-receptor type 22 RA Rheumatoid arthritis
RIG-I Retinoic acid-inducible gene I RLRs RIG-I-like receptors
RNA Ribonucleic acid
ROAS Norwegian National Registry of organ-specific autoimmune diseases SEAP Secreted embryonic alkaline phosphatase
SLE Systemic lupus erythematosus SNPs Single nucleotide polymorphisms ssRNA Single-stranded RNA
TBK1 TANK-binding kinase 1 TCR T cell receptor
TGF-β Transforming growth factor beta
Th T helper
TLR Toll-like receptor TNF Tumor necrosis factor Tregs Regulatory T cells
TRAF6 TNF receptor-associated factor 6
TRIF TIR-domain-containing adapter-inducing interferon-β T1D Type 1 diabetes
WES Whole-exome sequencing 21OH 21-hydroxylase
3βHSD2 3β-hydroxysteroid dehydrogenase type 2
List of publications
Paper I
Sigrid Aslaksen, Anette B. Wolff, Magnus D. Vigeland, Lars Breivik, Ying Sheng, Bergithe E. Oftedal, Haydee Artaza, Beate Skinningsrud, Dag E.
Undlien, Kaja K. Selmer, Eystein S. Husebye, Eirik Bratland (2019).
Identification and characterization of rare Toll-like receptor 3 variants in patients with autoimmune Addison's disease
Journal of Translational Autoimmunity, 1, 100005.
Paper II
Sigrid Aslaksen, Paal Methlie, Magnus D. Vigeland, Dag E. Jøssang, Anette B.
Wolff, Ying Sheng, Bergithe E. Oftedal, Beate Skinningsrud, Dag E. Undlien, Kaja K. Selmer, Eystein S. Husebye, Eirik Bratland (2019).
Coexistence of congenital adrenal hyperplasia and autoimmune Addison’s disease
Frontiers in Endocrinology, 10, 648.
Paper III
Alexander Hellesen * and Sigrid Aslaksen *, Lars Breivik, Ellen Christine Røyrvik, Øyvind Bruserud, Kine Edvardsen, Karl Albert Brokstad, Anette Susanne Bøe Wolff, Eystein S. Husebye, Eirik Bratland (2020).
Circulating 21-hydroxylase-specific CD8 + T cells in autoimmune Addison’s disease are predominantly restricted by HLA-A2 and HLA-C7 molecules.
Submitted to Science Immunology
*Both authors contributed equally
Introduction
Protection against invasive pathogens is fully dependent on the immune system, a highly evolved network of specialized cells and molecules. This host defense system is remarkably effective as serious persistent infections are quite rare. Nonetheless, there are some limitations in that severe infections occasionally may occur, and also, in that the immune response sometimes fails to discriminate self from non-self, causing autoimmunity.
The development of autoimmunity recapitulates the same immune responses used to fight off infections and results in destruction of the host’s own cells and tissues.
Autoimmune disorders may affect almost all organs, whereby the clinical phenotype and severity are dependent on which tissue is being attacked. In autoimmune Addison’s disease (AAD), the immune system is directed against the hormone-producing cells of the adrenal cortex, leading to insufficient production of vital steroids. Patients with Addison’s disease therefore depend upon lifelong replacement therapy [1].
In the following, a brief introduction of the main components of the immune system is given, followed by a description of immunological tolerance and autoimmunity with emphasis on AAD.
The innate and adaptive immune system
To establish an infection, the pathogen must break through the first line of defense consisting of the skin and mucous barriers that secretes chemical compounds restricting the adherence and growth of microbes. Pathogens that manage to overcome these barriers will meet the two further lines of defense, the innate and adaptive immune systems [2].
The components constituting the innate immune response may be found in all advanced living organisms and has been refined for a longer period throughout evolution than the adaptive immune system, which is believed to have arisen with the vertebrate lineage [3, 4]. The innate immune system mounts the same immediate response each
time it gets exposed to a certain pathogen, whereas the adaptive immune response will improve on every exposure to the same pathogen, making it antigen-specific.
The cellular components of innate immunity
The innate immune response is largely dependent on phagocytosis by mononuclear and polymorphonuclear phagocytes. The mononuclear phagocytes include monocytes, dendritic cells (DCs), and macrophages. DCs are professional antigen-presenting cells (APCs) that can activate immune cells of the adaptive immunity and are found in most tissues including lymphoid organs, skin, lung, and kidneys [3, 5]. Macrophages are specialized cells that engulf and destroy any foreign invader, as well as clearing up debris from dead cells, and are located within the parenchyma of major organs throughout the body. They also serve as APCs to activate adaptive immune responses.
Upon recognition of a pathogen, macrophages release chemotactic cytokines to attract other phagocytes, in particular polymorphonuclear phagocytes, to the site of infection.
The polymorphonuclear myeloid cells (also known as granulocytes) include neutrophils and eosinophils, releasing reactive oxygen species to kill bacteria and parasites, in addition to basophils and mast cells, the main mediators of inflammation and allergic responses [2]. Another important cell population of innate immunity is the innate lymphoid cells (ILCs). ILCs respond quickly to pathogens by producing various cytokines and are especially abundant at mucosal surfaces [6]. ILCs include the natural killer (NK) cells that can kill infected and malignant cells, in addition to produce diverse cytokines such as interferon gamma (IFNγ) and tumor necrosis factor alpha (TNFα) [7]. To further control the spreading of microorganisms, the complement system is activated, involving acute phase proteins in the blood that mediate opsonization and lysis of the pathogen, in addition to enhancing the inflammatory response [8].
To recognize pathogens, the innate immune system uses pattern recognition receptors (PRRs) that are specialized to detect small molecules deriving from viruses, fungi, bacteria, and protozoa, known as pathogen-associated molecular patterns (PAMPs) [9].
Binding of PAMPs to PRRs will activate numerous different signaling pathways that produce cytokines promoting the host response to infection [10, 11].
Innate immune receptors
The existence of PAMPs has been known for decades, but the PRRs recognizing them were introduced more recently by Charles Janeway in 1989 [12]. Since then, several classes of PRRs have been discovered including Toll-like receptors (TLRs), nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs), retinoic acid-inducible gene I (RIG-I)-like receptors (RLRs), C-type lectin receptors (CLRs) and AIM2-like receptors (ALRs). These receptors are categorized by their structure and specificity, in addition to their tissue-specific expression and cellular localization [11]. The best characterized PRRs are the TLRs, and at present ten functional human TLRs have been defined. TLR1, TLR2, TLR4, TLR5, TLR6, and probably TLR10 are all expressed on the cell surface, detecting extracellular components of bacterial and fungal cell walls, such as lipopolysaccharides (LPS), di– and triacyl lipopeptides, peptidoglycan (PGN) and flagellin. The remaining TLR3, TLR7, TLR8, and TLR9 are localized within endosomes, where they detect foreign nucleic acids such as double- stranded ribonucleic acid (dsRNA), single-stranded RNA (ssRNA) and unmethylated CpG deoxyribonucleic acid (DNA), respectively. These intracellular TLRs are believed to function primarily as viral sensors (Figure 1) [9, 11].
Activation of intracellular TLR signaling is induced by binding of the ligand to the leucine-rich repeats of the receptors’ ectodomain. For all TLRs except TLR3, this leads to activation of the adaptor protein myeloid differentiation primary response 88 (MyD88). MyD88 then interacts with interleukin (IL)-1R-associated kinase-4 (IRAK4), which phosphorylates IRAK1 and IRAK2 that, in turn, activate TNF receptor-associated factor 6 (TRAF6) [9, 13]. TRAF6 is further involved in activation of the mitogen-activated kinase (MAPK) pathway, and the two transcription factors nuclear factor-kappa B (NF-kB) and IFN regulatory factor 7 (IRF7), leading to the production of antiviral and proinflammatory cytokines like TNF, IL-12 and IL-6, and type I and III IFNs [14-16].
Upon TLR3 signaling, MyD88 is substituted with the adaptor protein TIR-domain- containing adapter-inducing IFN-β (TRIF), which interacts with both TRAF3 and TRAF6. TRAF3 links TANK-binding kinase 1 (TBK1) with inhibitor of NF-kB kinase subunit epsilon (IKKε), which together phosphorylate IRF3, leading to the production of type I and III IFNs. TRAF6, as mentioned earlier, induces transcription of proinflammatory cytokines and IFNs [14]. [17]
Figure 1. TLR signaling pathways. The extracellular TLRs (TLR1, TLR2, TLR4, TLR5, TLR6, and probably TLR10) are activated by molecules deriving from bacterial and fungal cell walls including lipopolysaccharides (LPS), di- and triacyl lipopeptides, peptidoglycan (PGN) and flagellin. The intracellular TLR3, TLR7, TLR8, and TLR9 are activated by nucleic acids within endosomes. Binding of the ligand to the receptors’ ectodomain induces activation of the adaptor proteins MyD88 or TRIF. MyD88 recruits IRAK4 to phosphorylate IRAK1 and IRAK2, which further phosphorylate TRAF6, leading to activation of the IKK complex (IKKγ, IKKα, and IKKβ). This complex activates NF-kB, allowing it to translocate into the nucleus and initiate the production of pro-inflammatory cytokines. The other adaptor protein TRIF interacts with both TRAF3 and TRAF6. TRAF3 links TBK1 with IKKε, which promotes phosphorylation of IRF3 and IRF7, leading to the production of type I and III IFNs.
The TLRs generally form homodimers upon activation, whereas TLR2 also functions as a heterodimer with TLR1 and TLR6. Figure modified from Wang et al [17].
Apart from the TLRs, RLRs also sense viral RNA products, not within endosomes, but in the cytosol of infected cells. RIG-I and melanoma differentiation-associated gene 5 (MDA5) are two well-defined RLRs, and activate NF-kB and IFN production through the same signaling pathways utilized by the TLRs [18].
Newly synthesized type I IFNs (α and β) and type III IFN (λ) will engage their cognate receptors to activate the Janus kinase/signal transducers and activators of transcription (JAK/STAT) signaling pathway. The result is the expression of IFN-stimulated genes (ISGs), which induce an antiviral state involving RNA degradation, inhibition of protein synthesis, upregulation of major histocompatibility complex (MHC) class I (described later), chemokine secretion, and apoptosis [19-22].
The adaptive immune system
While the innate immunity mediates a rapid response to infectious agents, the range of innate sensors recognizing them is limited. Consequently, an adaptive and more specific recognition system has evolved, providing a broader repertoire for detecting constantly mutating microbes. Also, after the initial exposure, the adaptive immune response develops memory cells with the capacity of mediating a rapid and enhanced response upon re-exposure to the same pathogen [23].
Cells of the adaptive immune system include thymus-derived T lymphocytes and bone marrow-derived B lymphocytes. During their development, they acquire T cell receptors (TCRs) and B cell receptors (BCRs), respectively, which enable them to recognize a wide range of antigens. The receptors are generated by random combinations of gene segments that are both variable and constant, giving rise to a wide T and B cell repertoire [2]. Following their development in the thymus and bone marrow, T and B cells migrate to secondary lymphoid organs for activation and proliferation. Activation is initiated upon binding of antigens by TCRs and BCRs, presented by professional APCs, in particular DCs, that have migrated to secondary lymphoid organs. The lymphocytes then migrate to the sites of infection and work together with the innate immunity to eliminate the pathogen [23]. There are two types of adaptive immune responses: the humoral response involving antigen-specific
antibodies produced by plasma cells (terminally differentiated B cells), and the cell- mediated response mediated by T cells, which destroys infected cells [24].
T cells
T cells can be divided into two main subsets based on their expression of either cluster of differentiation (CD) 8 or CD4 [25]. CD8+ T cells interact with antigen peptides presented by MHC class I molecules expressed by all nucleated cells, whereas CD4+ T cells interact with antigens presented by MHC class II molecules mainly restricted to APCs. The MHC proteins are encoded by the human leukocyte antigen (HLA) system including class I genes (HLA-A, HLA-B, and HLA-C) and class II genes (HLA-DR, HLA-DQ, and HLA-DP) [23]. In addition to antigen stimuli (signal 1), co-stimulatory signaling is required for T cell activation, involving the interaction of CD80/CD86 on APCs with CD28 on T cells (signal 2), as well as inflammatory cytokines (such as IL- 12 or IFNα/β) (signal 3) (Figure 2). The co-stimulatory signal is finely tuned by inhibitory receptors including cytotoxic T lymphocyte-associated protein 4 (CTLA-4) and programmed cell death protein 1 (PD1) [26].
Following these stimulating events, the production of IL-2 is induced and CD8+ T and CD4+ T cells differentiate into functionally distinct effector populations. Naïve CD4+ T cells, also known as T helper (Th) cells, develop into either Th1, Th2, or Th17 depending on the surrounding cytokine milieu mediated by innate immune cells. IL-12 results in differentiation toward Th1 cells, which support cell-mediated responses in which they secrete IFNγ and IL-2 to activate macrophages and CD8+ T cell development (Figure 2). IL-4 induces differentiation towards Th2 cells that stimulate humoral and allergic responses by releasing IL-4, IL-5, IL-10, and IL-13 [23, 24, 27, 28]. Transforming growth factor-beta (TGF-β) and IL-6 promote the development of Th17 cells, which secrete IL-17 important for protection against extracellular fungi and bacteria. Notably, Th17 cells are known to be implicated in the disease progression of several autoimmune disorders such as rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), and psoriasis [29, 30]. In addition to Th17 cell development, TGF-β, accompanied by IL-2, induces differentiation towards regulatory T cells
(Tregs), critical for regulating T-cell responses through anti-inflammatory cytokines, such as IL-10 [31].
Activation of naïve CD8+ T cells results in the conversion to cytotoxic T cells that are essential for defeating pathogens (mostly viruses) in the cytosol of infected cells. Once they are activated, they produce IFNγ to stimulate their cytotoxicity and motility, enabling them to efficiently colonize sites of inflammation [32]. Importantly, IFNγ production is associated with several autoimmune diseases including SLE and AAD [33-35]. Following migration to the site of inflammation, cytotoxic T cells recognize antigens presented by MHC class I molecules on the surface of target cells and kill them by mediating the coordinated action of perforin and granzymes. Perforin Figure 2. A simplified schematic of T cell activation. Extracellular antigens are taken up by an APC, in this figure a DC, and presented to CD4+ T cells via class II MHC molecules and CD8+ T cells via class I MHC molecules (Signal 1). For proper T cell activation, co- stimulatory signals are required, involving the interaction of CD80/CD86 on the APC with CD28 on the T cells (Signal 2), as well as inflammatory cytokines produced by the APC (Signal 3). Release of cytokine IL-12 will push the CD4+ T cell into the Th1 type, which, in turn, will secrete cytokines, such as IL-2, to support T cell proliferation and differentiation.
Figure produced using Medical Servier Art.
generates pores in the membrane of the target cell, allowing granzymes to enter and activate a caspase cascade. Ligation of Fas-ligand on the cytotoxic T cell to Fas receptors on the target cell can also induce apoptosis, although this mechanism functions mainly to delete activated Fas-expressing lymphocytes after an infection has been cleared, to maintain lymphocyte homeostasis [23, 36].
B cells and antibodies
The adaptive humoral response is mediated by antibody-producing plasma cells developed from B cells. In the bone marrow, B cells pass through several differentiation stages involving rearrangements of the genes encoding antibody, or immunoglobulin (Ig), heavy and light chains to become IgD and IgM-expressing cells.
This stage does not require contact with exogenous antigens and is therefore called antigen-independent B cell development [23]. In the next phase, B cells migrate to secondary lymphoid organs to finalize their Ig repertoire. This development stage is mainly dependent on the ability of B cells to function as APCs. The IgM and IgG receptors capture antigens and internalize them for further presentation by MHC class II molecules on the B cell surface. When the B cell then interacts with a T cell, in particular a Th cell specific for the MHC class II/antigen, the T cell starts producing cytokines to promote B cell proliferation. Subsequent interaction between CD40L on T cells and CD40 on B cells promotes differentiation of B cells into short-lived - or memory plasma cells, as well as Ig class switching from IgM and IgD to IgG, IgA or IgE, depending on the surrounding cytokine milieu [23, 27].
Immunological tolerance
Cells of the adaptive immunity are shaped to recognize and neutralize an infinite number of pathogens. Inevitably, some of their receptors may also react to host components. T and B cells must therefore undergo immunological tolerance enabling them to discriminate between self and non-self. There are two mechanisms by which immunological tolerance occurs: i) during lymphocyte maturation in the primary lymphoid organs (central tolerance); ii) during lymphocyte-antigen interactions in the
secondary lymphoid organs (peripheral tolerance). If these mechanisms fail, lymphocytes may start to attack cells of our own body resulting in autoimmunity [37, 38].
Central tolerance
The process of central B cell tolerance is thought to be less complex than the one for T cells. Immature B cells that display self-reactive receptors in the bone marrow either undergo apoptosis or get the chance to rearrange their receptors’ specificity to become less self-reactive, resulting in survival [38]. On the other hand, T cells are required to pass through two tolerance mechanisms: positive- and negative selection. Positive selection represents the first step where only double-positive T cells (CD4+CD8+) showing a defined affinity to self MHC molecules survive. Those without any affinity, constituting 90% of all double-positive cells, undergo a death pathway called “death by neglect” [39]. Following positive selection, double-positive T cells differentiate into single-positive cells (either CD4+ or CD8+) before subjected to negative selection in the thymic medulla. Here, they encounter a variety of self-antigens by interacting with medullary stromal cells including medullary thymic epithelial cells (mTECs) [40]. By utilizing promiscuous gene expression, mTECs and other APCs in the thymus express and present numerous peripheral tissue-specific antigens [41, 42]. These antigens represent diverse tissues such as the lung, heart, stomach, and kidney, and their expression is regulated by the key transcription factor “autoimmune regulator” (AIRE).
AIRE is therefore a critical regulator of central tolerance, and deficiency causes autoimmune destruction of several, mostly endocrine, organs manifesting as the syndrome autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), also known as autoimmune polyendocrine syndrome type-1 (APS-1) [43, 44]. Besides this well-established role for AIRE in negative selection, there is increasing evidence that AIRE also promotes thymic Treg development [45]. A recent study of Aire-deficient mice showed that AIRE promotes the generation of Foxp3+CD4+ Tregs in the perinatal period [46]. Given this influence, it has been speculated that Treg defects caused by AIRE deficiency might contribute to autoimmunity [45]. However, this hypothesis requires further testing.
After surviving both positive and negative selection, the remaining CD4+ and CD8+ T cells expressing TCRs without high affinity for self-peptides, migrate to the periphery and secondary lymphoid organs where they have the potential to interact with their cognate antigens presented by APCs [37].
Peripheral tolerance
Although central tolerance is thought to be efficient, some self-reactive lymphocytes manage to escape deletion in the primary lymphoid organs. Thus, peripheral tolerance has evolved to serve as a second elimination process in the periphery. This mechanism is based upon the principle that any escaping self-reactive lymphocyte would need to find a cooperating partner for sufficient co-stimulation and activation. In the absence of such cooperation, autoreactive lymphocytes enter an inactive state, known as anergy, which remains even if they re-encounter the same antigen under appropriate stimuli.
For autoreactive B cells, anergy is induced when they interact with self-antigens without finding a cooperating Th cell. In this condition, B cells are outcompeted by other B cell clones and become excluded from entering the lymphoid follicles to receive survival factors [37, 47].
When self-reactive T cells escape negative selection, peripheral tolerance is promoted by at least three distinct mechanisms: i) anergy; ii) deletion by activation-induced cell death; iii) immune suppression by Tregs [37]. Anergy is induced, as previously mentioned, when the TCR encounters antigens in the absence of co-stimulation. It can also be induced by binding of CD80/CD86 on APCs to CTLA-4 on T cells, which transmits an inhibitory signal [48, 49]. The second mechanism, deletion by activation- induced cell death, is induced when Fas (CD95) and FasL (CD95L) on T cells interact after repeated stimulation of the TCR [50]. The last peripheral tolerance process is mediated by Tregs, which suppress T cell responses either in a contact-dependent manner or by producing immunosuppressive cytokines such as TGF-β or IL-10 [51].
Tregsare therefore important for maintaining self-tolerance, which is underlined by previous research showing that decrease in their numbers and function is associated with several autoimmune disorders such as SLE, multiple sclerosis (MS), type 1
diabetes (T1D), APS-1, RA, and immune dysregulation, polyendocrinopathy and enteropathy, X-linked (IPEX) syndrome [52-54].
Autoimmunity
From the brief overview of immunological tolerance above, it appears that autoimmunity is a result of defective elimination of self-reactive lymphocytes. Once some of these autoreactive lymphocytes are activated, multiple parts of the immune system may get involved in attacking host tissue, leading to either systemic or organ- specific autoimmune disorders. So far, more than 80 autoimmune disorders have been described, and their incidence and prevalence are still rising [55]. Some of them seem to have shared, yet complex, genetic bases and etiology pathways as they often tend to co-occur within individuals and families [56]. However, they may only develop after exposure to certain environmental factors, making them challenging to study and fully understand.
Genetic factors
Several genetic risk factors for autoimmunity have been identified, but their precise contribution is often hard to specify. Nonetheless, a few autoimmune disorders are induced by single gene alterations, making them easier to study and diagnose. One classic example is the unique disorder APS-1 caused by both autosomal recessive mutations and dominant-negative mutations in AIRE [44, 57, 58]. Another example is the autoimmune lymphoproliferative syndrome, also known as the Canale-Smith syndrome, caused by impairment of Fas-induced apoptosis, due to mutations in the FAS gene or other genes involved in apoptosis [59]. The IPEX syndrome is also a monogenic autoimmune disease, caused by mutations in forkhead box P3 (FOXP3) that regulates the development and suppressive activity of Tregs [60, 61].
Apart from these disorders, most autoimmune diseases are polygenic, resulting from multiple susceptibility variants in many genes. The best-characterized genetic component so far is mapped to the HLA locus encoding class I or II MHC molecules.
Several associations between polymorphisms in the HLA complex and autoimmune diseases have been described. For instance, the haplotypes DRB1*03:01-DQA1*05:01- DQB1*02:01 and DRB1*04:04-DQA1*03:01-DQB1*03:02 (DR3-DQ2 and DR4- DQ8) are shown to confer risk to both celiac disease (CD), T1D, autoimmune thyroid disease (AITD), and AAD, explaining why these diseases tend to co-occur within individuals [43, 62, 63]. CD is nonetheless considered one of the most evident HLA- associated diseases as it occurs selectively in individuals expressing the DQ2 or DQ8 haplotypes [64, 65]. In conjunction, specific HLA alleles seem also to be involved in drug-induced autoimmunity by abacavir, an anti-retroviral medication used to treat human immunodeficiency virus (HIV). In treated individuals expressing the HLA- B*57:01, abacavir seems to alter the binding cleft of the HLA class I molecule, increasing self-epitope binding. This leads to autoreactive T cell responses and abacavir hypersensitivity reactions [66, 67].
The role of certain HLA alleles in the pathogenesis of autoimmunity is most probably related to their antigen-presenting capacity; either by enhancing peptide-presentation in the periphery, leading to increased activation of T cells or by an insufficient presentation of self-antigens in the thymus, resulting in more self-reactive T cells or fewer Tregs [68, 69]. Apart from the HLA locus, several other autoimmune susceptibility genes have been identified, encoding cytokines, costimulatory molecules, molecules involved in promoting apoptosis, members of cytokine- or antigen-signaling cascades, and molecules that clear antigen or antigen-antibody complexes [70]. None of these genetic risk factors, however, can alone induce autoimmunity as it is the overall genetic background of the host that determines the probability of disease development [70].
Autoimmune susceptibility genes can also increase immunoreactivity, but they may not necessarily lead to an improved immune response against pathogens. In fact, individuals with autoimmune disorders show increased susceptibility to infections, and manifestations of autoimmunity may only occur after exposure to certain pathogens [71, 72]. This is mirrored by the high degree of genetic overlap between autoimmune disorders and monogenic primary immunodeficiency diseases (PIDs). Variations in the
PID genes including AIRE, STATs, FOXP3, LRBA (lipopolysaccharide responsive beige-like anchor), and CTLA4 are reported to be involved in the immune dysregulation of several autoimmune diseases including APS-1, IPEX, T1D, RA, vitiligo and AITD [73, 74]. Patients with PID caused by mutations in these genes are also prone to develop autoimmune diseases. Therefore, a major reason for autoimmunity could be the lack of clearance of microbial antigens, implying the impact of environmental triggers.
A useful tool for studying the relative contribution of genetic and environmental factors in autoimmune diseases is twin studies. In CD and AAD, the concordance rates among monozygotic twins are 75-86% [75, 76] and 71% [77] respectively, whereas RA and systemic sclerosis have less concordance [78], suggesting higher involvement of environmental factors. Moreover, a true twin concordance rate may only develop after several years, meaning that monozygotic twins have different disease onsets, also suggesting the involvement of non-genetic factors [78]. Another compelling evidence for an environmental contribution is the genotype-independent geographic variation of β-cell autoimmunity involved in T1D [79].
Environmental factors
The well-known environmental factors associated with autoimmunity include infections, vaccines, hormones, chemical exposures, drugs, diet, and cigarette smoking [80-82]. Among these, the most compelling and studied one is infection caused by viruses, bacteria, and other pathogens [83]. Several mechanisms of how infections induce autoimmune responses have been proposed. The pathogen may carry epitopes that are similar to the host’s antigens, a phenomenon termed “molecular mimicry”.
Lymphocytes that react to such epitopes may then also react to self-antigens, resulting in tissue damage and activation of other parts of the immune system. Another mechanism is “epitope spreading”, whereby the host tissue is damaged by the immune response to a persistent infection, and/or by necrosis caused by the pathogen itself.
Self-antigens are then released, leading to local activation of APCs and presentation of self-peptides that further activate autoreactive lymphocytes. Activation and stimulation of these lymphocytes may also be indirect or non-specific, caused by inflammation
present during an infection, a mechanism called “bystander activation” involving
“cryptic antigens” [84]. In contrast to dominant antigens, cryptic antigens are usually invisible to the immune system as they appear in low concentrations. Upon inflammation, however, their concentration can increase due to elevated production of proteases and differential processing of released self-peptides by APCs [85, 86]. This creates the opportunity for autoreactive lymphocytes to develop.
The role of infection underlying autoimmune responses has especially been assigned to T1D, MS, and autoimmune liver disease or autoimmune hepatitis (AIH) [87-89].
Concerning T1D, previous research has demonstrated the presence of enteroviruses (EV) in pancreatic islets from patients [89]. Epidemiological studies have also reported more EV infections in T1D patients than in controls [89, 90]. Regarding MS, several epidemiological and immunological studies indicate an association with Epstein-Barr virus (EBV) [91-96]. However, the results remain controversial; some researchers provide evidence for the expression of EBV in B cells in inflamed post-mortem brain tissue [91], while others do not detect EBV-positive cells, or at least in very few patients only [97-100]. In patients with AIH, previous research demonstrates the presence of antibodies against hepatitis C virus (HCV), in addition to a type of autoantibodies called liver/kidney microsomal antibody type 1 (LKM-1) in patients with chronic HCV infection [87]. Notably, many of the infections associated with autoimmune disorders are opportunistic, primarily affecting individuals who cannot efficiently eradicate the virus, leading to persistent proinflammatory conditions.
Impaired immunity and increased susceptibility to infections have also been demonstrated in patients with AAD in terms of reduced NK cell cytotoxicity and poor responses to IFNs in-vitro [101, 102]. These findings are mirrored by studies reporting that AAD patients have a higher intake of antibiotics and antivirals, both before and after the diagnosis, compared to the general population [103, 104].
Autoimmune Addison’s disease
While Addison’s disease during the last century was predominantly caused by tuberculous infiltration of the adrenal glands, the autoimmune form currently dominates in developed countries where it is responsible for approximately 80% of the cases in adult patients [83, 105, 106]. AAD is a classic organ-specific autoimmune disease characterized by an immune-mediated attack on the hormone-producing cells in the adrenal cortex, while the adrenal medulla remains intact. Destruction of the adrenal cortex leads to deficiency of the vital steroids cortisol and aldosterone, and patients with AAD therefore depend upon lifelong supplementation therapy [1, 107].
The adrenal cortex
The adrenal gland is a hormone-producing organ, composed of two cell layers with distinct embryological origins. The inner adrenal layer (medulla) arises from the ectoderm lineage and produces catecholamines (such as epinephrine and norepinephrine), whereas the outer adrenal cortex derives from intermediate mesoderm and converts cholesterol into various bioactive steroid hormones (Figure 3 and 4). The adrenal cortex consists of three separate zones, whereby each produces different steroids mediating stress response and regulation of blood pressure. The zona glomerulosa forms the outermost layer and produces the mineralocorticoid aldosterone, important for the maintenance of blood pressure. Zona glomerulosa is under the control of the renin-angiotensin system, which is regulated by blood pressure and changes in sodium and potassium levels. The next cell layer constitutes zona fasciculata that synthesizes glucocorticoids like cortisol under stimulation by the hypothalamic- pituitary-adrenal (HPA) axis. Cortisol induces gluconeogenesis and anti-inflammatory processes during stress. The innermost zona reticularis secrete androstenedione and dehydroepiandrosterone (DHEA), which can further be converted to sex hormones in peripheral tissues [108, 109].
Figure 3. Adrenal anatomy. The adrenal glands are located on top of the kidneys and consist of the outer adrenal cortex and the inner adrenal medulla. The adrenal cortex is subdivided into three distinct layers: Zona glomerulosa, producing mineralocorticoids, zona fasciculata, secreting glucocorticoids, and zona reticularis, producing androgens. The inner medulla synthesizes catecholamines including epinephrine and norepinephrine. Figure produced using Medical Servier Art.
Figure 4. Steroidogenesis in the adrenal cortex. Cholesterol is converted to aldosterone, cortisol, and androgens through different pathways that require the specific enzymes cholesterol side-chain cleavage enzyme (CYP11A), 17α-hydroxylase (CYP17), 3β- hydroxysteroid dehydrogenase type 2 (3βHSD2), 21-hydroxylase (CYP21), 11β-hydroxylase (CYP11B1), and aldosterone synthase (CYP11B2).
Clinical characteristics
AAD is a relatively rare disorder with a prevalence in European populations ranging from 93-220 cases per million [105, 110-114]. Like other autoimmune diseases, AAD affects more women (female: male ratio 1.5-3.5:1) and can occur at any age, although it generally affects individuals between 30 to 50 years of age [105]. Patients have insufficient production of the adrenal steroids due to the autoimmune attack of the adrenal cortex. The pathological changes can develop slowly over many years before the clinical manifestations appear, which include fatigue, nausea, dizziness, weight loss, skin hyperpigmentation, and salt-craving [105, 107, 115]. Hyperpigmentation is due to the absence of the cortisol-mediated negative feedback mechanism of the HPA axis response, leading to increased production of adrenocorticotropic hormone (ACTH) and pro-opiomelanocortin (precursor of ACTH), which enhances stimulation of the melanocortin 1 receptor of skin melanocytes [116, 117]. Salt-craving is caused by aldosterone deficiency, which impairs the kidney’s ability to reabsorb salt, causing low blood pressure. This increases the activity of renin to activate the protein hormone angiotensin that, in turn, tries to stimulate the release of aldosterone by the adrenals [115, 118]. Elevated ACTH levels and plasma renin activity are therefore important clinical characteristics of AAD. To correct the insufficient levels of cortisol and aldosterone, patients are treated with hydrocortisone and fludrocortisone, respectively [115].
The dominant target of the adrenal autoantibodies seems to be the steroid cytochrome P450 21-hydroxylase (21OH) and has ever since its discovery in 1992 been the important diagnostic marker of AAD [119]. 21OH is located in the smooth endoplasmic reticulum (ER) of adrenocortical cells and are involved in the synthesis of cortisol and aldosterone (Figure 4) [120]. Autoantibodies against 21OH are reported to be present in >95% of patients at diagnosis in developed countries [121]
and 86% of all Norwegian patients [122]. This high frequency, however, tends to decline with a disease duration greater than 20 years. Furthermore, a previous study reported that 21OH autoantibody-positive patients with other autoimmune endocrinopathies or first-degree relatives of patients with organ-specific autoimmune
diseases have a cumulative risk for AAD of around 50% [123]. Thus, presence of these antibodies may predict AAD development. Among the general population, the frequency of 21OH autoantibodies is approximately 0.5% [124].
AAD may not only occur in isolation. Around 50% of AAD patients also suffer from other autoimmune entities such as T1D, AITD, premature ovarian insufficiency (POI), CD, and autoimmune gastritis [124]. The co-existence of at least two of the three endocrine diseases T1D, AITD, and AAD is typically referred to as APS-2 [43]. AAD is also frequently present in the unique disorder APS-1 (prevalence 1:90 000 in Norway) defined clinically by the presence of two out the three manifestations: AAD, hypoparathyroidism, and mucocutaneous candidiasis [125]. Whereas APS-1 is a monogenic disease, AAD (and APS-2) seem to have a more complex etiology involving multiple genes and unknown environmental factors [126].
The genetic basis of AAD
Investigating the genetic basis of complex diseases, such as AAD, is challenging due to several factors. The genotype of a patient does not necessarily correlate with the phenotype; phenocopies may occur, meaning that environmental factors could induce an AAD phenotype in a patient without known disease susceptible variants that matches the phenotype of a patient that does carry such variants. Additionally, AAD is a genetically heterogeneous disease in which variants in any one of several susceptibility genes may cause the same phenotype, and multiple of these may be necessary to produce it (polygenic inheritance). Lastly, AAD show incomplete penetrance, which means that certain individuals who inherit susceptibility alleles might not develop the disease [105].
Despite all these complicating factors, research on monozygotic twins and familial clustering of AAD have revealed the existence of highly penetrant susceptible alleles [105, 127, 128]. The best characterized seem to be located within the HLA class II haplotypes HLA-DR3-DQ2 and HLA-DR4-DQ8. Polymorphisms in these haplotypes enable the presentation of a different repertoire of self-antigens in the peptide-binding pocket. Notably, individuals carrying the heterozygous combination DR3-DQ2/DR4-
DQ8 are at higher risk than homozygotes carrying either one of the haplotypes (Figure 5) [113, 129]. This may be due to the formation of heterodimers by the trans-encoded DQ molecules, increasing the functional diversity of the MHC complex [130]. HLA class I genes have also been associated with AAD, more specifically those that are part of the conserved allele combination DR3-B8 (Figure 5) [62, 63]. These alleles are further part of the ancestral haplotype defined as 8.1 (AH 8.1) containing the core HLA alleles HLA-A*01:01, -C*07:01, -B*08:01, -DRB1*03:01, -DQA1*05:01 and - DQB1*02:01 [131]. Apart from AAD, AH 8.1 has been associated with a wide range of immune-mediated diseases including T1D, CD, SLE, AIH, common variable immunodeficiency (CVID), and IgA deficiency [132, 133].
Single nucleotide polymorphisms (SNPs) in genes regulating T and B cell activation and differentiation are also associated with increased risk of AAD development (summarized in Table 1). Among these, SNPs in CTLA4 and PTPN22 (protein tyrosine phosphatase non-receptor type 22) appear frequently as genetic risk variants (also in other autoimmune disorders), underlining the importance of fine-tuning T cell signaling to avoid development of autoimmunity [134-137]. CTLA-4 downregulates T cell signaling by binding to CD80/86 on APCs, and gene polymorphisms impairing this mechanism may lead to reduced inhibition of T cell responses [138, 139]. PTPN22
Figure 5. Combination of HLA alleles conferring high risk of AAD development. Figure produced using Genome Decoration Page (NCBI).
inhibits T cell activation by dephosphorylating key signaling molecules downstream of the TCR. This process is dependent on the binding of PTPN22 to C-terminal Src kinase tyrosine kinase, and SNPs (in particular c.1858C>T, p.Arg620Trp) interrupting this physical interaction cause lack of T cell inhibition [138, 140, 141]. Other susceptibility variants are located in the genes encoding BACH2 (BTB domain and CNC homolog 2), important for B cell differentiation and generation of Tregs [129, 142]; NALP1 (NACHT leucine-rich-repeat protein 1), promoting inflammasome assembly and inflammatory cytokine release [143, 144]; and CLEC16A (C-Type Lectin Domain Containing 16A), presumably regulating autophagy and NK cell function [145].
Recently, variants in AIRE were also found to be associated with AAD, at least in the Swedish patient cohort [146]. This finding has yet to be replicated in other populations, although many AAD risk loci may differ between studies of different cohorts.
Table 1. Susceptible genes contributing to AAD, outside the MHC region.
Gene Function References
CTLA-4 Inhibits T cell responses [134], [137]
PTPN22 Dephosphorylates signaling molecules downstream of the TCR, important for T cell inhibition
[135], [136]
BACH2 Promotes B cell differentiation and generation of Tregs
[129], [142]
NALP1 Promotes inflammasome assembly and inflammatory cytokine release
[143], [144]
CLEC16A Regulates autophagy and NK cell function [145]
AIRE Master regulator of negative selection in the thymus [146]
None of the risk genes described so far, except for HLA, seem to have a major impact on an individual’s genetic susceptibility to develop AAD; they only explain a small fraction of familial clustering. An explanation for this missing heritability problem is that many of the available genotyping assays focus on SNPs present in more than 1- 5% of the patient population [147]. However, the genetic architecture of AAD seems to be complex, involving multiple variants. An important step towards solving this missing heritability problem is therefore to search for rare and low-frequency variants with high impact. This strategy could illuminate the genetic variation in AAD, which is important for better prevention, diagnosis, and treatment of the disease.
Immunopathogenesis
The specific triggers underlying the progression of AAD from one phase to another are still not clear, but there have been considerable advances in our understanding of the pathogenesis over the past few decades.
Histologic studies of adrenal glands from patients with AAD show a significant mononuclear cell infiltration and atrophy of the adrenal cortex, in addition to some fibrosis, while the medulla is spared [148, 149]. These data suggest that during the active phase of the disease there is an extensive destruction and loss of all three cell layers of the adrenal cortex mediated by lymphocytes, plasma cells, and macrophages [150]. Since AAD is primarily associated with class II HLA alleles, it seems likely that initiation of adrenocortical destruction depends upon activation of CD4+ T cells that, in turn, license autoreactive immune cell responses. As there is a potentially unlimited supply of self-antigens, this autoreactive immune attack may continue until the target organ is destroyed and replaced by fibrous tissue.
Irreversible damage of both zona glomerulosa and zona fasciculata is what causes the first clinical signs of adrenal insufficiency [105, 150]. Notably, this might not be until 90% of all adrenocortical cells are destroyed [151]. Before overt, symptomatic adrenal failure, however, autoantibodies against 21OH are typically present. Whether these autoantibodies are involved in disease progression is yet not clear, although it appears that they have no inhibiting effect on the 21OH enzyme in vivo [107, 150, 152]. This
is not that surprising since 21OH is an intracellular enzyme. Also, a previous study demonstrated that transplacental passage of 21OH autoantibodies from a mother with AAD was not sufficient to cause autoimmune adrenocortical destruction in the child [153]. Thus, the main mediators of the autoimmune destruction seem to be cytotoxic T cells, although the specific mechanisms remain unclear.
The first indication of T cell proliferation in response to adrenal proteins was provided by Freeman and Weetman, using peripheral blood mononuclear cells (PBMCs) from AAD patients [154]. However, no individual adrenal antigens were identified. Husebye et al. in contrast, used BALB/c and SJL mice, immunized with recombinant 21OH, to demonstrate proliferation of lymph node cells to a specific 21OH-derived epitope, 21OH342-361 [155]. A few years later, Bratland et al. showed T cell proliferation and IFN-γ production in response to the same epitope (21OH342-361),using PBMCs from AAD patients [33]. Since then, epitope mapping has suggested two specific 21OH epitopes targeted by circulating autoreactive CD8+ T cells. The first epitope, published by Rottembourg et al. [34], was mapped to the HLA-B8 restricted sequence EPLARLEL (21OH431-438) after performing an ELISPOT assay to detect IFN-γ producing T cells in response to 21OH peptides of 20 amino acids length.
Subsequently, using MHC-peptide multimer technology, two responder patients were shown to have higher frequencies of EPLARLEL-specific CD8+ T cells than a HLA- B8-positive control. A few years later, a second epitope was revealed by Dawoodji et al. [156]. This one was mapped to the HLA-A2 restricted sequence LLNATIAEV (21OH342-350). T cells specific for this sequence were reported to have the ability to lyse cells pulsed with the LLNATIAEV-peptide, by granzyme B secretion. Together, these studies suggest a central role for 21OH-specific T cells in the pathogenesis of AAD and open up new avenues for research into the features of these cells.
The major and unanswered question of the immunopathogenesis is what causes the lymphocytic infiltration of the adrenal cortex. This issue is challenging to address because all histochemical analyses of adrenal glands from AAD patients are post- mortem due to practical and ethical reasons. Also, the experimental animal models developed so far, have shown limited relevance to the human AAD [105]. Most studies
on the pathogenesis of AAD have therefore been restricted to peripheral blood and serum. Interestingly, evidence of lymphocytic infiltration in the adrenal gland has not only been limited to AAD patients. A previous study on autopsy cases of young and old individuals revealed that approximately 7% of the subjects less than 50 years had mononuclear cell infiltration within the adrenal cortex, and this frequency increased with age [157]. These infiltrations are most likely not significantly pathogenic, as the prevalence of AAD is far more rare, but it might indicate that focal infiltration of the adrenals does occur. In such circumstances, a healthy adrenal gland is producing a sufficient amount of cortisol that would ensure an anti-inflammatory milieu, protecting the cortex from an autoimmune attack [158]. This protective and immunosuppressive effect of cortisol could then help explain why AAD is a rare condition compared to other autoimmune endocrine disorders [105]. In that case, a key step in the pathogenesis of AAD could therefore be a reduction of cortisol at sites of infiltration (e.g. due to adrenocortical damage) which may reduce tolerance to adrenal proteins and lead to activation of lymphocytes. As a result, steroid production decreases along with increased infiltration, resulting in a vicious cycle of adrenocortical cell damage and steroidogenesis impairment. Together, these events may trigger an autoreactivity to adrenal antigens, eventually leading to a complete adrenal failure [105]. The exact mechanisms underlying this selective autoimmune attack and reduced cortisol production in AAD patients remain elusive, but it has been postulated that virus infection of the adrenal gland is a potent trigger [83].
Viruses as triggers of the autoreactive immunity
The coordinated attack of the adrenocortical cells, involving CD4+ T cells, CD8+ T cells and antibody-producing B cells, driven mainly by the self-antigen 21OH, is highly reminiscent of the specific immune response against viruses. In this context, we could speculate that viral infection of the adrenal gland leads to necrosis of adrenocortical cells, resulting in the release of self-antigens and augmentation of lymphocytic infiltration, causing inflammation of the adrenal cortex. This may further lead to uptake of self-antigens and/or viral peptides by APCs. Since 21OH and other steroid enzymes are among the most abundant proteins in the adrenal cortex, they are likely to be
presented as peptides on MHC molecules. Release, uptake, and presentation of self- antigens (epitope spreading) can further activate autoreactive T and B cells, and thereby eventually lead to autoimmune adrenalitis (Figure 6) [83].
This hypothetical model for AAD development is supported by previous studies showing that the adrenals are permissive to a variety of viral agents including cytomegalovirus (CMV), EBV, herpes simplex virus types 1 and 2 (HSV-1/HSV-2), HCV and polyomaviruses [83]. Most of these viruses cause opportunistic infections and usually do not lead to autoimmunity, suggesting the requirement of additional host genetic errors and immunodeficiencies. Indeed, this seems to be the case for AAD Figure 6. A potential viral pathogenic mechanism in AAD. A simplified model of how virus infection of the adrenal cortex may lead to activation of autoreactive immune cells and adrenal insufficiency. First, persistent viral infection causes necrosis of adrenocortical cells and subsequently release, uptake, and presentation of self-antigens by APCs. Secondly, autoreactive T cells get activated by self-antigen presentation within local draining lymph nodes, which, in turn, license the activation of macrophages (M) and autoreactive B cells.
Further necrosis and apoptosis of adrenocortical cells cause decreased cortisol production, increasing the lymphocyte infiltration in the adrenal cortex, eventually leading to AAD. Figure produced using Medical Servier Art.
patients, as previously mentioned [101-103, 159]. Viral pathogenic mechanisms are also suggested for other complex immune-mediated diseases like asthma, T1D, and Crohn’s disease [160]. Unlike these diseases, however, no specific pathogens have been identified for AAD, highlighting the need for research on viral exposure in affected patients. In this way, we may gain a better understanding of the pathogenesis, and thereby design therapies to protect susceptible individuals with inappropriate antiviral responses.
Aims
The overall objective of this project was to identify and functionally characterize genetic risk factors involved in AAD pathogenesis. To fulfill this aim, we used a two- sided strategy: First, we validated and characterized specific mutations revealed by whole-exome sequencing (WES) of 142 AAD patients. Since AAD is a rare condition with a complex genetic basis, we set out to search for and characterize rare genetic variants (allele frequency <1% in the general population) predicted to- or previously demonstrated to be damaging, which could make up for the missing heritability problem in AAD. Secondly, we aimed to elucidate genetic and immunological factors that influence disease progression from one phase to the other. For this, we focused on HLA-specific immunological determinants for autoreactive T cells, as these cells are thought to be the main driver of adrenal gland destruction in AAD.
The specific aims were:
I. Validate and characterize rare TLR3 variants identified in several unrelated patients.
II. Investigate clinical aspects of a patient carrying a rare homozygous nonsense mutation in the HSD3B2 gene.
III. Characterize autoreactive CD8+ T cells from AAD patients to identify HLA- class I restricted immunodominant peptides of 21OH.
Methodology
In this section, the most central parts of the methodology used in the thesis will be discussed. A detailed description of the materials and methods can be found in the separate papers.
Patient and control material
Through the Norwegian Registry of organ-specific autoimmune diseases (ROAS), we have access to serum, plasma, EDTA blood, DNA and, PBMCs from almost 1000 AAD patients. All have signed a written informed consent approved by the Health Region West Ethics committee (2018/1417, 146/96-47·96). Samples from healthy controls were collected from blood donors provided by the blood bank at Haukeland University Hospital. All experiments were conducted following the declaration of Helsinki.
Reporter gene assay (paper I)
To determine the functional impact of the TLR3 variants revealed by WES, we studied the response to polyinosinic:polycytidylic acid (poly(I:C)) stimulation in cells transfected with constructs encoding the wild type or variant TLR3 proteins. This was achieved using Hek Dual Null cells, which are stably transfected with a secreted embryonic alkaline phosphatase (SEAP) reporter gene induced by NF-kB. As the endogenous TLR3 gene is specifically knocked out, all expression and activity of the receptor are therefore dependent on transfected TLR3. Following transfection, we assessed TLR3 activity by monitoring NF-kB-induced SEAP production after poly(I:C) stimulation, which was measured by spectrophotometry in the cell supernatants after adding QUANTI-Blue solution.
MHC class I/peptide multimers (paper III)
The MHC multimer technology is a highly useful tool to detect and quantify antigen- specific T cells from PBMCs or other sources/tissues. Recombinant MHC molecules are loaded with peptides of antigen proteins, which can be recognized by cytotoxic T cells. Several of these MHC-peptide complexes are attached to a fluorescent backbone that can be detected and quantified by flow cytometry [161].
In paper III, we used MHC class I/peptide multimers (dextramers and streptamers) containing either HLA-A2-, HLA-B8- or HLA-C7-bound peptides. For details about the HLA-A2/B8-dextramers, see paper III. The MHC I-streptamers consisted of an HLA-C7-peptide complex attached to strep-tags, which are short peptides with binding affinity to a streptavidin molecule, named Strep-Tactin. Strep-Tactin molecules are fluorescently labeled, enabling quantification by flow cytometry.
HLA-C typing (paper III)
Since the HLA types of all the patients included in paper III were already estimated by allelic imputation by a genome-wide association study (GWAS) [162], we needed to find a quick and reliable way to HLA-type healthy controls. We decided to perform a PCR using sequence-specific primers for HLA-C7, published by Bunce et al. [163].
PBMCs and DNA from 51 healthy controls were then collected for HLA-typing, whereby 30% turned out to be positive for the HLA-C*0701 allele and thereby included in the study.
